Direct road mortality is one of the main mechanisms in which transport networks affect wildlife, with potential negative effects on population abundance and persistence. Although road-kill data are relatively easy to collect and databases are available worldwide, there are many challenges in analyzing road-kill data to allow better inferences and decisions regarding road mitigation. Important research questions might be explored using road-kill data: how many animals are killed on the road, which proportion of the population is being removed by road mortality, and where is this impact concentrated. To estimate how many animals are killed, it is necessary to correct for imperfect detection due to searcher inefficiency and carcass persistence. Different sampling schemes and analytical approaches are being applied to correct the estimates, and this is already the standard practice in other research areas. New approaches should also be tested, such as the use of hierarchical modeling to consider the spatial and temporal variation in these sources of errors. By correcting mortality estimates, results could be compared among roads, times and observer teams. In some contexts, an important challenge is to assess the repercussion of road fatalities for population size and persistence and little effort has been done to compare the outcomes of different available approaches to attain that goal. Ultimately, we want to to know where to intervene to reduce fatalities and many different methods are available to analyze road-kill spatial patterns and identify road-kill hotspots. However, care must be taken when comparing roads with varying lethality, since road-kill hotspots only inform where absolute road-kill numbers are higher, but not necessarily where are the populations under higher risk. The development and use of proper analytical approaches for each question and for each type of data is fundamental for a better comprehension of wildlife mortality and mitigation planning on transport networks and my aim on this talk is to present some of our latest findings and discuss some research gaps.